← Back to Model Beat
4Models·2d ago

Qwen-Image-2.0-RL Technical Report

Alibaba researchers have released a technical report for Qwen-Image-2.0-RL, a new pipeline designed to refine the model's visual output and instruction adherence. The process utilizes reinforcement learning from human feedback and on-policy distillation to enhance the performance of the underlying diffusion model.

ModelsQwen Image

Covered by 1 source

  • AarXiv CS.AI↗Yixian Xu, Kaiyuan Gao, Yuxiang Chen, Yilei Chen, Zecheng Tang, Zihao Liu, Zikai Zhou, Deqing Li, Hao Meng, Kuan Cao, Jiahao Li, Jie Zhang, Liang Peng, Lihan Jiang, Ningyuan Tang, Shengming Yin, Tianhe Wu, Xiaoyue Chen, Yan Shu, Yanran Zhang, Yi Wang, Yu Wu, Yujia Wu, Zekai Zhang, Zhendong Wang, Xiao Xu, Kun Yan, Chenfei Wu2d ago

Related stories

ModelsAlibaba's new open Qwen image model aims for more natural-looking results - the-decoder.comDec 31ModelsQwen-Image-Agent: Bridging the Context Gap in Real-World Image GenerationJun 26ModelsIs the Chinese Version of Nano Banana Available? Qwen-Image-2.0 Shines: Effortlessly Handles 1K Long Texts, No More Awkwardness in Chinese Image GenerationFeb 11ModelsAlibaba's Overnight Bombshell: Qwen - Image - 2.0 Tested, Excels in Outfit Changes, Group Photos, and Clear WritingFeb 11